Genetic Programming Technique Research Articles

An enhanced stability evaluation system for entry-type excavations: Utilizing a hybrid bagging-SVM model, GP and kriging techniques

In underground mining, especially in entry-type excavations, the instability of surrounding rock structures can lead to incalculable losses. As a crucial tool for stability analysis in entry-type excavations, the critical span graph must be updated to meet more stringent engineering requirements. Given this, this study introduces the support vector machine (SVM), along with multiple ensemble (bagging, adaptive boosting, and stacking) and optimization (Harris hawks optimization (HHO), cuckoo search (CS)) techniques, to overcome the limitations of the traditional methods. The analysis indicates that the hybrid model combining SVM, bagging, and CS strategies has a good prediction performance, and its test accuracy reaches 0.86. Furthermore, the partition scheme of the critical span graph is adjusted based on the CS-BSVM model and 399 cases. Compared with previous empirical or semi-empirical methods, the new model overcomes the interference of subjective factors and possesses higher interpretability. Since relying solely on one technology cannot ensure prediction credibility, this study further introduces genetic programming (GP) and kriging interpolation techniques. The explicit expressions derived through GP can offer the stability probability value, and the kriging technique can provide interpolated definitions for two new subclasses. Finally, a prediction platform is developed based on the above three approaches, which can rapidly provide engineering feedback.

Comparison of boosting and genetic programming techniques for prediction of tensile strain capacity of Engineered Cementitious Composites (ECC)

Plain concrete is weak against tension and has low Tensile Strain Capacity (TSC) which significantly affects its long-term performance. To overcome this issue, Engineered Cementitious Composites (ECC) were developed by incorporating polymer fibres in the cement matrix which increases ductility and provides higher TSC than plain concrete and they have emerged as a viable alternative to brittle plain concrete. This study is conducted in an attempt to develop empirical prediction models for TSC prediction of ECC without requiring extensive experimental procedures. For this purpose, two evolutionary programming techniques known as Multi Expression Programming (MEP), Gene Expression Programming (GEP) along with two boosting-based techniques: AdaBoost and Extreme Gradient Boosting (XGB) were developed using data collected from published literature. The gathered dataset had seven input parameters including water-to-binder ratio, sand, fibre content, cement, fly ash, superplasticizer, and age etc. and only one output parameter i.e., TSC. The error assessment of developed models was done using correlation coefficient, Mean Absolute Error (MAE), and Objective Function (OF) etc. and the error comparison showed that XGB has the highest accuracy having the least OF value of 0.081 as compared to 0.11 of AdaBoost, 0.13 of GEP, and 0.16 of MEP. Shapley additive analysis was conducted on the XGB model since it proved to be the most accurate, and the results highlighted that fibre content, age, and water-to-binder ratio are the most important features to predict TSC of ECC.

On the application of symbolic regression in the energy sector: Estimation of combined cycle power plant electrical power output using genetic programming algorithm

This paper focuses on the estimation of electrical power output (Pe) in a combined cycle power plant (CCPP) using ambient temperature (AT), vacuum in the condenser (V), ambient pressure (AP), and relative humidity (RH). The study stresses accurate estimation for better CCPP performance and energy efficiency through responsive control to changing conditions. The novelty lies in applying genetic programming (GP) on a publicly available dataset to generate Symbolic Expressions (SEs) for high-accuracy Pe. To address the challenge of numerous GP hyperparameters, a random hyperparameter values search method (RHVS) is introduced to find optimal combinations, resulting in SEs with higher accuracy. SEs are created with varying input variables, and their performance is evaluated using multiple metrics (coefficient of determination (R2), mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), Kling–Gupta Efficiency (KGE), and Bland–Altman (B-A) analysis). A key innovation involves combining the best SEs through an Averaging ensemble (AE), leading to a robust estimation accuracy. Notably, the AE YVE−2 achieves the highest (Pe) accuracy, including R2=0.9368, MAE=3.3378, MSE=18.4800, RMSE=4.2985, MAPE=0.7354%, and KGE=0.9479. The investigation highlights AT as the most influential variable, underscoring the importance of choosing inputs aligned with physical processes. This paper’s outlined procedure, combining GP, hyperparameter optimization, and ensemble techniques, offers an efficient method for estimating Pe in CCPP. It promises simplicity and effectiveness in real-world applications. B-A analysis proves valuable for SE selection, enhancing the proposed methodology.

Survey on Genetic Programming and Machine Learning Techniques for Heuristic Design in Job Shop Scheduling

Job shop scheduling is a process of optimising the use of limited resources to improve the production efficiency. Job shop scheduling has a wide range of applications such as order picking in the warehouse and vaccine delivery scheduling under a pandemic. In real-world applications, the production environment is often complex due to dynamic events such as job arrivals over time and machine breakdown. Scheduling heuristics, e.g., dispatching rules, have been popularly used to prioritise the candidates such as machines in manufacturing to make good schedules efficiently. Genetic programming, has shown its superiority in learning scheduling heuristics for job shop scheduling automatically due to its flexible representation. This survey firstly provides comprehensive discussions of recent designs of genetic programming algorithms on different types of job shop scheduling. In addition, we notice that in the recent years, a range of machine learning techniques such as feature selection and multitask learning, have been adapted to improve the effectiveness and efficiency of scheduling heuristic design with genetic programming. However, there is no survey to discuss the strengths and weaknesses of these recent approaches. To fill this gap, this paper provides a comprehensive survey on genetic programming and machine learning techniques on automatic scheduling heuristic design for job shop scheduling. In addition, current issues and challenges are discussed to identify promising areas for automatic scheduling heuristic design in the future.

Application of multigene genetic programming and water evaporation optimization technique for modeling and optimization of removal of heavy metals from ash pond water using cyanobacterial consortium

Abstract Heavy metals such as Lead(II), Nickel(II), Manganese(II), Cadmium(II), Chromium(VI), etc., are leached from the coal ash of thermal power plants. These metals contaminate ash pond water and subsequently contaminate groundwater. Phycoremediation using microalgae/cyanobacteria is an emerging technology for removal of heavy metals. The present study aims at phycoremediation of the said heavy metals from ash pond water using cyanobacterial consortium of Limnococcus limneticus and Leptolyngbya subtilis followed by the development of an accurate data-driven Multigene Genetic Programing (MGGP) approach for modeling and optimization of the process. The developed model was used to obtain a correlation between the average removal of metals and biomass production with all input factors such as the initial metal concentrations, pH, and days of incubation. To maximize metal removal and biomass production, the Water Evaporation Optimization (WEO) technique was applied to determine optimal values of input parameters. The application of WEO for the optimization of the phycoremediation process is the first of its kind and here lies the novelty of the study.

Data Augmentation for Genetic Programming-Driven Late Merging of HOG and Uniform LBP Features for Texture Classification

The categorization of texture images requires the identification and extraction of meaningful keypoints, a crucial step in ensuring the precise representation of textured images. The literature has introduced numerous descriptors in order to detect and capture both local and global texture characteristics. These descriptors vary in their effectiveness depending on the specific application. However, it is generally accepted that they complement each other by compensating for their strengths and weaknesses. Therefore, many approaches focused on combining different types of image descriptors generating robust features that are invariant to image transformations. These solutions mostly focused on one way to deal with information and faced more problems as they need human intervention and a large set of data. The acknowledged benefits of combining multiple types of descriptors are accompanied by challenges specially. These challenges arise due to differences in properties, such as locality and sparsity, as well as the heterogeneity exhibited by generated features. To address this issue comprehensively, we present an approach that benefits from genetic programming techniques to generate and combine two distinct texture classifiers. It incorporates histograms of oriented gradients and local binary patterns descriptors, which capture different textures. To effectively fuse the results of both classifiers, the proposed approach employs a late fusion process along with data augmentation methods while using a limited amount of training data. To evaluate the performance of the proposed approach in texture image classification tasks, we have conducted extensive experiments on six challenging datasets encompassing various variations. We have also investigated its performance in a cross dataset problem where the model has been trained on instances of a dataset before being tested on samples of another dataset. The obtained results clearly demonstrate that the proposed approach surpasses other relevant low-level approaches, as well as existing GP-based and CNN methods specifically designed for describing and classifying textures. Thanks to its ability to simultaneously leverage multiple descriptors, the suggested solution shows a high potential for real-world applications, particularly in handling various image changes with robustness.

Exploring conduction mechanisms in chalcogenide thin films: an experimental and soft computing approach with ANN and GP techniques

The article consists of two main sections. In the first part, an experimental study was conducted to investigate the AC conduction of Se83Bi17 (SB), Se83Bi17Ge5 (SBG), and Se83Bi17Te5 (SBT) in thin film forms. Thermal evaporation was utilized to produce films with different thicknesses. The measurements were performed at temperatures ranging from 293 to 393K, film thicknesses ranging from 230.2 to 543.9 nm, and frequencies ranging from 0.1 to 1000 kHz. The AC data were analyzed using the frequency power law. The values of the frequency exponent and their temperature dependence confirmed that the conduction process in the thin films is governed by the hopping mechanism. Additionally, the impact of introducing Te and Ge on the studied properties was also investigated in this study. The second section of the article focuses on estimating the AC conduction of Se83Bi17 (SB), Se83Bi17Ge5 (SBG), and Se83Bi17Te5 (SBT) films using Artificial Neural Networks (ANNs) and Genetic Programming (GP). The experimental data was employed as inputs for training the ANNs and the GP models. The simulation and prediction results obtained from both models were compared with the experimental data and reported. Equations were developed to describe the behavior of the experimental data. The study concluded that the ANN approach is more suitable for simulating and forecasting AC conductivity.

A PREDICTIVE EQUATION FOR WAVE SETUP THROUGH THE USE OF GENETIC PROGRAMMING

As the sea level rises, coastal flooding is predicted to increase worldwide. One of the processes contributing to estimating coastal flooding is wave setup (combined with swash), which accounts for the effect of breaking waves on water levels. Defined as the superelevation of the mean water level due to breaking waves (Longuet-Higgins and Stewart, 1964), wave setup prediction has been a research topic of interest for decades. However, empirical predictors still show considerable scatter, making this component the one with the largest uncertainty when estimating flooding levels. At the same time, more data has become available, opening the possibility of using data-driven models, such as machine learning, a powerful tool that has already been applied to predict a variety of coastal processes. Our work aims to develop a new, robust and reliable wave setup predictor through the use of an evolutionary-based genetic programming technique. To develop the algorithm, we use a dataset compiled by Stockdon et al. (2006), containing 491 measurements from 10 field experiments representing different beach and wave conditions.

Developing deterministic and probabilistic prediction models to evaluate high-temperature performance of modified bitumens

This study aims to develop deterministic and probabilistic prediction models for the multiple stress creep and recovery (MSCR) test. For this purpose, crumb rubber, polyphosphoric acid, and styrene–butadiene–styrene bitumen modifiers have been used with different dosages to modify high-temperature performance of PG 58–28 and PG 64–22 base bitumens. The MSCR test has been performed at different temperatures. Deterministic models are developed by the multi-gene genetic programming technique for each modifier individually, and the non-recoverable creep compliance (Jnr) and percent recovery (R) parameters are predicted. The accuracy of deterministic models is suitable and the performance of R models has been better than Jnr models. Furthermore, a comprehensive probabilistic model has been developed by using the logistic regression technique to predict different traffic levels. The accuracy of the probabilistic model is 0.85. The sensitivity analysis has been performed on this model and the effect of changes in the modifier dosage and temperature on the traffic levels have been investigated. Results show that using the probabilistic model, it is possible to find a range of modifier’s dosage in which the traffic level is desired.

Estimating Unconfined Compression Strength of Fly Ash and Cement Stabilised Clayey Soil using Artificial intelligence Techniques

The study aims to develop proposed predictive formulas for determining the unconfined compression strength (UCS) of fly ash and cement stabilised clayey soil based on Multi-Gene Genetic Programming (MGGP) and Artificial Neural Network (ANN) techniques. Thirteen parameters, including the soil characteristics, the binder types, the binder contents, the curing period, the mixing method, and the fly ash characteristics, such as calcium oxide (CaO) content, CaO/SiO2 ratio, loss of ignition, were considered as the independent variables in the model. The results show that the selected optimal ANN and MGGP models can predict the target values with high correlation coefficients (R-value approximately 0.994 and 0.973, respectively), and low errors. The performances of the MGGP and ANN models were compared based on statistical parameters and several external criteria. The study finds that both models show their generalisation capabilities with robust, powerful, and accurate prediction ability; however, the ANN model slightly outperforms the MGGP model. The proposed predictive equations formulated from the selected optimal MGGP and ANN models could help engineers and consultants to choose the suitable binder and a reasonable amount of fly ash in the pre-planning and pre-design period.

Advanced Genetic Programming vs. State-of-the-Art AutoML in Imbalanced Binary Classification

The objective of this article is to provide a comparative analysis of two novel genetic programming (GP) techniques, differentiable Cartesian genetic programming for artificial neural networks (DCGPANN) and geometric semantic genetic programming (GSGP), with state-of-the-art automated machine learning (AutoML) tools, namely Auto-Keras, Auto-PyTorch and Auto-Sklearn. While all these techniques are compared to several baseline algorithms upon their introduction, research still lacks direct comparisons between them, especially of the GP approaches with state-of-the-art AutoML. This study intends to fill this gap in order to analyze the true potential of GP for AutoML. The performances of the different tools are assessed by applying them to 20 benchmark datasets of the imbalanced binary classification field, thus an area that is a frequent and challenging problem. The tools are compared across the four categories average performance, maximum performance, standard deviation within performance, and generalization ability, whereby the metrics F1-score, G-mean, and AUC are used for evaluation. The analysis finds that the GP techniques, while unable to completely outperform state-of-the-art AutoML, are indeed already a very competitive alternative. Therefore, these advanced GP tools prove that they are able to provide a new and promising approach for practitioners developing machine learning (ML) models. Doi: 10.28991/ESJ-2023-07-04-021 Full Text: PDF

Application of multi-gene genetic programming technique for modeling and optimization of phycoremediation of Cr(VI) from wastewater

BackgroundRemoval of Cr(VI) from wastewater is essential as it is potentially toxic and carcinogenic in nature. Bioremediation of heavy metals using microalgae is a novel technique and has several advantages such as microalgae remove metals in an environmentally friendly and economic manner. The present study deals with modeling and optimization of the phycoremediation of Cr(VI) from synthetic wastewater. The initial concentration of Cr(VI), initial pH, and inoculum size were considered as input factors, and the percentage removal of Cr(VI) was chosen as a response.ResultsAn accurate data-driven genetic programming model was developed with the experimental data of other scientists to find a relation between the percentage removal of Cr(VI) and all input parameters. To maximize the removal of Cr(VI), the grey wolf optimization technique was applied to determine the optimal values of input parameters.ConclusionThese optimum input parameters are difficult to get through experimentation using the trial-and-error method. The established modelling and optimization technique is generic and can be applied to any other experimental study.Graphical

Predicting liquefaction-induced lateral spreading by using the multigene genetic programming (MGGP), multilayer perceptron (MLP), and random forest (RF) techniques

Predicting liquefaction-induced lateral spreading by using the multigene genetic programming (MGGP), multilayer perceptron (MLP), and random forest (RF) techniques

Global optimization and monotonicity in entropy production of weak drivings.

Knowing if an optimal solution is local or global has always been a hard question to answer in more sophisticated situations of optimization problems. In this paper, for finite-time and weak isothermal driving processes, we show the existence of a global optimal protocol for the entropy production. We prove this by showing its convexity as a functional in the derivative of the protocol. This property also proves its monotonicity in such a context, which leads to the satisfaction of the second law of thermodynamics. In the end, we exemplify that the analytical technique of the Euler-Lagrange equationapplied to overdamped Brownian motion delivers the global optimal protocol, by comparing it with the results of the global optimization technique of genetic programming.

Axial Gap Motor Shape Optimization Using Multi-Objective Based Genetic Programming Technique

Axial Gap Motor Shape Optimization Using Multi-Objective Based Genetic Programming Technique

The Electrical Conductivity of Ionic Liquids: Numerical and Analytical Machine Learning Approaches

In this paper, we incorporate experimental measurements from high-quality databases to construct a machine learning model that is capable of reproducing and predicting the properties of ionic liquids, such as electrical conductivity. Empirical relations traditionally determine the electrical conductivity with the temperature as the main component, and investigations only focus on specific ionic liquids every time. In addition to this, our proposed method takes into account environmental conditions, such as temperature and pressure, and supports generalization by further considering the liquid atomic weight in the prediction procedure. The electrical conductivity parameter is extracted through both numerical machine learning methods and symbolic regression, which provides an analytical equation with the aid of genetic programming techniques. The suggested platform is capable of providing either a fast, numerical prediction mechanism or an analytical expression, both purely data-driven, that can be generalized and exploited in similar property prediction projects, overcoming expensive experimental procedures and computationally intensive molecular simulations.

Can We Estimate Air‐Sea Flux of Biological O2 From Total Dissolved Oxygen?

AbstractIn this study, we compare mechanistic and empirical approaches to reconstruct the air‐sea flux of biological oxygen () by parameterizing the physical oxygen saturation anomaly (ΔO2[phy]) in order to separate the biological contribution from total oxygen. The first approach matches ΔO2[phy] to the monthly climatology of the argon saturation anomaly from a global ocean circulation model's output. The second approach derives ΔO2[phy] from an iterative mass balance model forced by satellite‐based physical drivers of ΔO2[phy] prior to the sampling day by assuming that air‐sea interactions are the dominant factors driving the surface ΔO2[phy]. The final approach leverages the machine‐learning technique of Genetic Programming (GP) to search for the functional relationship between ΔO2[phy] and biophysicochemical parameters. We compile simultaneous measurements of O2/Ar and O2 concentration from 14 cruises to train the GP algorithm and test the validity and applicability of our modeled ΔO2[phy] and . Among the approaches, the GP approach, which incorporates ship‐based measurements and historical records of physical parameters from the reanalysis products, provides the most robust predictions (R2 = 0.74 for ΔO2[phy] and 0.72 for ; RMSE = 1.4% for ΔO2[phy] and 7.1 mmol O2 m−2 d−1 for ). We use the empirical formulation derived from GP approach to reconstruct regional, inter‐annual, and decadal variability of based on historical oxygen records. Overall, our study represents a first attempt at deriving from snapshot measurements of oxygen, thereby paving the way toward using historical O2 data and a rapidly growing number of O2 measurements on autonomous platforms for independent insight into the biological pump.

A Coupled Genetic Programming Monte Carlo Simulation–Based Model for Cost Overrun Prediction of Thermal Power Plant Projects

Globally, power projects are prone to cost overrun projects. Within the body of knowledge, previous studies have paid less attention to predicting the cost overruns to assist contingency cost planning. Particularly, in thermal power plant projects (TPPPs), the enormous risks involved in their delivery undermine the accuracy of cost overrun prediction. To prevent cost overrun in thermal power plant projects, these risks need to be accounted for by employing sophisticated cost overrun prediction techniques. This study aims to develop a hybrid predictive-probabilistic-based model (HPPM) that integrates a genetic programming technique with Monte Carlo simulation (MCS). The HPPM was proposed based on the data collected from TPPPs in Bangladesh. Also, the sensitivity of the HPPM was examined to identify the critical risks in cost overruns simulation. The simulation outcomes show that 40.48% of a project’s initial estimated budget was the most probable to cost overrun, while the maximum cost overrun will not exceed 75% with 90% confidence. Practically, the analysis will sensitize project managers to emphasize thermal plants’ budget accuracy not only at the initial project delivery phase but throughout the project life cycle. Theoretically, the HPPM could be employed for cost overrun prediction in other types of power plant projects.

Estimating microscale DE parameters of brittle adhesive joints using genetic expression programming

Particle-based model has strength and flexibility in modelling the microstructures of adhesives and interface in adhesive joints. In this work, a procedure with genetic expression programming (GEP) technique to calibrate the microscale parameters of discrete element (DE) model was proposed for brittle adhesives. Two categories of adhesive properties, the bulk property of thick adhesive and interlaminar-like property of thin adhesive, were discussed. For the bulk property, three target properties of adhesives, i.e. tensile strength, peak strain, secant modulus, were set as the reproduced features. 300 sets of adjustable microscale parameters were produced to run the numerical tests and generate datasets. GEP was then employed to find regression formulas for predicting the target properties as a function of the microscale parameters. For the interlaminar-like property, fracture energies of the cohesive failure of thin adhesives were approximated. A similar procedure of combined DE modelling and GEP was performed to find the regression models to estimate the fracture energy. The developed regression formulas can cover a general range of brittle adhesives. Loctite EA 9497 adhesive was selected to perform a series of lab tests, of which the results were subsequently used to examine the applicability of the DE model with calibrated parameters. The numerical results exhibit good agreements with testing data and observation.

A Metaheuristic Optimization Algorithm for Solving Higher-Order Boundary Value Problems

An effective metaheuristic algorithm to solve the higher-order boundary value problems, called a genetic programming technique is presented. In this paper, a genetic programming algorithm, which depends on the syntax tree representation, is employed to obtain the analytical solutions of higher- order differential equations with the boundary conditions. The proposed algorithm can be produce an exact or approximate solution when the classical methods lead to unsatisfactory results. To illustrate the efficiency and accuracy of the designed algorithm, several examples are tested. Finally, the obtained results are compared with the existing methods such as the homotopy analysis method, the B-Spline collocation method and the differential transform method.